It seems you have nothing to contribute to this topic. Besides OP has already "left the building". If you have nothing better to do on a Saturday evening just study carefully the thread in question and the claims you made there.
I still keep reading, but it seems the conversation has shifted from IC recommendation to what the heck is doppler distortion
I wish you luck on your dac build. To me ES9033 seems the way to go. In the same fashion as higher-grade ESS chips it has I2C access to a pile of registers so you will also gain experience about that as well. The downside is that these registers are probably poorly documented but that is the story also with other ESS (and AKM) chips.
Its a thing speakers do. Doesn't matter, the point was that typical audio FFTs omit phase information which is commonly justified by a belief some people have that humans are insensitive to phase. Its a belief that humans can't hear any difference simply due to phase, at least not above a few hundred Hz. The Purifi files certainly debunk that phase-insensitivity myth. Humans can be shown to be phase insensitive or not depending on test waveforms used. Music contains all sorts of waveform patterns. Sometimes phase information matters a lot in music reproduction. So why do we keep ignoring phase in FFT measurements?
Look, we are just scratching the surface of what is wrong with how we measure audio. We don't need to go deeper into it here other than that I would like you to understand that dac manufacturer published specs are not the whole story. Don't be dazzled by a few impressive numbers and let that guide your dac choices. Don't know if you will learn that first time out or even learn it ever. Some people just don't seem to be able to handle the complexity of the physical world. They want things to be simple, clear-cut, easy to understand 'facts'. If only it were that easy.
Last edited:
Thank you! I only had a short look at the documentation, but it seems to be not too bad. I keep you updated on the progress in the other thread
Can you please lay out an other measurement technique that will not miss anything? I mean just the measurement - skip the interpretation vs. hearing...
Doesnt have to be too detailed, just hint of the basic mechanisms...
//
With that kind of logic, why would I buy expensive Hifi Audio stuff? Not only are those companies making big claims about specs, but using a lot of adjectives too...
for example, this is just a random pick from a cable product:
These new xxxxxxxxxxxxxxx power cables use Direction-Controlled Long-Grain Copper (LGC) in a Semi-Solid Concentric Topology—a sophisticated physical arrangement in which strands are packed more tightly and never change position within the bundle. Direction-Controlled LGC minimizes distortion caused by grain boundaries that exist in any metal conductor, which in turn assures induced noise is dissipated and drained properly—as always, we direct noise where it will do the least harm to your system’s performance.
I'm not saying they are right or wrong, I could have picked up something else too. And I don't want to start a new discussion. I'm aiming for a "mid-tier" DIY project and as mentioned above well executed puts the project in the area of some cheaper DACs (which btw are well-regarded nowadays).
I don't understand why you make this assumption. I try to be as objective as I can making the decision to choose the DAC, listening to recommendations while doing my research.
Some just don't seem to be able to understand the complexity of simple things. It often comes down to simply making a choice, some hope, some luck and many hours tweaking. All the measurements aside some devices will never please the ear for whatever reason. I once made a small series of 5 DACs with a Philips/NXP all-in-one chip that performed so bad that after having tried all 5 of them I took a hammer and made them to powder.
Keeping stuff as simple as possible is often ridiculed but I think it is the way to go. This counts both for the circuits (many devices suffer from bad power supplies and almost none suffer from good power supplies) and operation of the device. Please just have a look at the elderly and also think if you yourself or your partner/children will be able to simply use the device correctly. If not then the device is unsuitable for normal use. After having made such "niche in a niche" devices I now know these are almost always one offs that end up in a drawer. Indication: if even the thought of reusing poses the need for a many hours remembering how the thing worked or should be programmed.
Last edited:
https://www.diyaudio.com/community/threads/placement-of-resistors-in-signal-path.384534/post-6988928
Those files are a just a marketing gimmick to lure gullible people. Rotating phase 90 degrees is not something commonly occurring in analog audio devices.
Please don't buy expensive gear based on advertising claims. A lot of it isn't very good, although there are a few gems.
Audio is difficult in the sense that its aimed at creating an illusion that works with the human auditory perceptual system. Video reproduction is analogous. The world is not RGB, but we know how to fool the visual system into seeing something credible enough. The auditory system so far is much less well understood. The specs measured at, for example, ASR, do not strongly correlate with auditory perception. Some people may want to believe the specs tell it all, but really they are being fooled by putting too much meaning into semi-meaningless numbers. Maybe you will see the truth of that someday. I don't get the feeling you understand it yet.
Being arrogant again? Why do you need to belittle OP just because he does not buy into your audiophile mumbo jumbo?
Sure. But why that approach can come closer to satisfying human perception is a complex subject. There can be many simple variations to try in a circuit to fine tune the sound to find what amounts to a complex interaction between components that is satisfying to the ear.
Last edited:
Not trying to be rude. Its like someone claims to have an objective view of reality, but doesn't yet understand that perception of reality is an illusory model created inside the brain. Its good enough to help increase chances of survival, but it is not objective truth. We know that very well in regard to vision. We know the perception of color is an illusion. There is nothing inherently 'yellow' about a school bus. The word yellow refers to perceptual experience in the brain. A hypothetical space alien that sees light as frequencies of EM radiation doesn't necessarily ever experience a sense of 'yellow.' Its like false coloring a map where yellow is used to encode height. There is nothing inherently yellow about height just like there is nothing inherently yellow about a school bus. Yellow describes how certain information is mentally experienced, that's all, nothing more.
And so it is that SINAD is to some extent useless (as are similar measurements). Sure its a rough estimate of something, but not exactly how we hear or exactly what we need to fool the auditory system like we fool the visual system in order to convincingly experience reproduction of sensed reality.
Last edited:
I didn't suggest anything.
If one cant come up with an alternative way of measurement one probably don't have the competence to diss the existing ones.
//
just a follow up on my design questions regarding the DAC.
Is it generally advised to use different BUCK/LDOs for the MCU and the DAC, although they share the same voltage level eg 3v3 and GND and sit on the same PCB?
Looking through some datasheets and ref designs a potential implementation could be to use a LT8619 as a buck with adjustable output or fixed 5 V, and then something like a LT3042 as the LDO.
The concept of designing a system with low complexity would apply here too, according to the information that LT is providing. the requirement of good documentation is reference designs would also be fulfilled.
Is it generally advised to use different BUCK/LDOs for the MCU and the DAC, although they share the same voltage level eg 3v3 and GND and sit on the same PCB?
Looking through some datasheets and ref designs a potential implementation could be to use a LT8619 as a buck with adjustable output or fixed 5 V, and then something like a LT3042 as the LDO.
The concept of designing a system with low complexity would apply here too, according to the information that LT is providing. the requirement of good documentation is reference designs would also be fulfilled.
Do you think it has to be exactly 90 degrees to be audible? Phase rotation is used all the time in radio to control the modulation index of the transmitter. Ever listen to a radio?
Also, phase shift occurs all time when systems have minimum phase properties, which all do. Transients get smeared out over time, and the effect can be audible. The problem is when humans rely on FFT measurement where phase information has been discarded. Some humans tend to believe the spectral lines tell all, but the lines can't tell all because information is missing.
May I ask if you guys use this thread to discuss the topic or related questions? There's already a lot of offtopic in here and I intend to gather information regarding my implementation rather than ideas that don't seem to correlate.
Your question was based on the premise that a single tell-all test should described by me when you know very well its more complicated than that. Its like asking someone, "have you stopped beating your wife?" Its the premise that is contrived by the person asking that is intended to setup the person being asked.
Sure. However, that should apply to you as well. Please don't ask questions if you don't want to hear the responses the question elicits.